Two original one-shot vector cues, one layered character scene, and one numeric-input field. The cues include editable SVG source, Lottie JSON, and complete static fallbacks. The scene shows how named SVG parts respond to several states. The field shows how a pure model can drive a browser or native renderer from the host app's number.
| Example | Use | Motion | Files |
|---|---|---|---|
| Bookmark saved | Small acknowledgment beside a host interface's successful save status | 5 frames at 30 fps (167 ms) | SVG source and still, Lottie JSON, color binder |
| Review draft | Transition from a processing frame to a draft that needs human review | 9 frames at 30 fps (300 ms) | SVG source and still, Lottie JSON |
| Character scene | Explore a five-state character and small pointer-linked depth scene | Interruptible state changes, quiet idle loop; optional six-second video render | Single layered SVG, browser controller, still poster, render script, MP4 test |
| Signal field | Explore a generic numeric input with no health or product meaning | Direct, reversible changes to separate tiles | Pure model, browser control, still SVG, React Native sample, optional Lottie/dotLottie companions |
| Native adapters | Reanimated versions of the kept cue and signal field, plus a Reduce Motion and lifecycle hook | UI-thread tweens and input-driven styles | KeptCue.jsx, SignalFieldReanimated.jsx, useMotionPolicy.js, pure policy test |
| Models | Paced cycle (breathing, pacing) and stage stepping for any ordered set | Level and cue events from elapsed time; pose from a continuous position | paced-cycle.js, stage-step.js, tests; native/usePacedCycle.js drives the cycle from one Reanimated frame clock |
These examples are visual cues. The host application owns the state transition, wording, accessibility announcement, and next action. Show the saved cue only after a real save succeeds; show the draft cue only when the draft is actually ready. A partial or failed operation should use its own state. Motion must never block the next action or imply that content was verified.
From the repository root:
python -m http.server 8000 --bind 127.0.0.1Open http://127.0.0.1:8000/preview/ for the one-shot cues, http://127.0.0.1:8000/examples/character-scene/ for the character, or http://127.0.0.1:8000/examples/signal-field/ for continuous input. The cue preview has Replay, Settled state, sample colors, and a Reduce Motion switch. It uses a local copy of lottie-web 5.13.0 with its own MIT notice. The examples use no CDN, account, or analytics.
The named groups in each source.svg are editable in a vector editor. tools/build_library.py defines matching geometry, colors, and keyframes; update both when changing the design, then regenerate the exports. The bookmark's raw Lottie uses placeholder colors. Bind it and its static SVG to the host theme's accent and line hex colors with tokens.mjs before display.
The character scene guide explains its single SVG source, five-state controller, still fallback, optional MP4 render, and reproduction commands. The scene page needs the local HTTP server because it loads that SVG as editable DOM. Its optional video test starts only when a viewer presses Play. The video is a framed, opaque sample; it cannot substitute for interactive states.
The signal field guide explains its numeric model, browser and React Native renderers, static fallback, and optional host-scrubbed Lottie exports. Its source and publication record names the factual Apple reference and the visual changes that keep this public example distinct. No private reference media was copied.
npm ci
npm run build # regenerates every export; outputs are byte-reproducible
npm test # structure, token, scene, model, policy and native bundle checksCI runs the same build and fails if any committed export differs, then runs the tests. The native check bundles the React Native sources with Expo's default Metro extensions and leaves React, React Native and Reanimated external. It proves syntax and imports, not native rendering.
To evaluate low-cost idle loops for an AI-generated mascot still, use the Generated mascot motion skill and its scripts in tools/mascot/: an image-to-video loop on a chroma frame, exported per platform (WebM with alpha, stacked-alpha MP4, HEVC-with-alpha on macOS) with a manifest, and a frame-swap Lottie built from aligned image edits. It includes stage-by-stage questions and a character card template. Its routes and costs record a measured comparison. Generation calls need your own provider keys and are billed by those providers.
For a new codebase, use the portable Create app motion skill. Copy that skill folder into your agent's skills directory, then ask it to adapt one real state or input in your app. It guides source inspection, route choice, original art, still and reduced motion behavior, and target-renderer checks. These examples are starting points, not universal interaction patterns. Browser playback and video decoding do not establish acceptance in a native renderer or on a device.
Verification level. The Lottie and dotLottie files are browser-verified only (lottie-web 5.13.0, dotlottie-web 0.80). The native adapters are source-verified: they bundle, and Reanimated 4.5.1's worklet plugin compiles them. None has yet run in a React Native development build or on a device. See VERIFICATION.md.
assets/**/*.svg,assets/**/*.lottie.json, andexamples/character-scene/scene.svg,scene-poster.png, andscene-loop.mp4: Creative Commons Attribution 4.0 International (CC BY 4.0), © 2026 KFLCHI001. See full CC BY 4.0 text. Credit: “Quiet Motion Library artwork by KFLCHI001, CC BY 4.0,” link to this repository and the license, and indicate changes when applicable.examples/signal-field/still.svg,signal-field.json, andsignal-field.lottie: CC BY 4.0, © 2026 KFLCHI001, with the same credit and change notice.tools/,skills/,preview/index.html,assets/bookmark-kept/tokens.mjs,examples/character-scene/HTML, JavaScript, Python and render requirements,examples/signal-field/HTML, JavaScript, JSX, Python and documentation,native/,models/,.github/,package.json, repository documentation, and configuration: MIT, © 2026 KFLCHI001. See full MIT text.preview/vendor/lottie.min.js: upstream MIT under its separate notice.
The asset manifest records provenance and intended meaning. No product logo, screenshot, private text, third-party source image, or downloaded community animation is included.